AAAI Publications, Twenty-First International Joint Conference on Artificial Intelligence

Font Size: 
Minimal Module Extraction from DL-Lite Ontologies using QBF Solvers
Roman Kontchakov, Luca Pulina, Ulrike Sattler, Thomas Schneider, Petra Selmer, Frank Wolter, Michael Zakharyaschev

Last modified: 2009-06-25

Abstract


We present a formal framework for (minimal) module extraction based on an abstract notion of inseparability w.r.t. a signature between ontologies. Two instances of this framework are discussed in detail for DL-Lite ontologies: concept inseparability, when ontologies imply the same complex concept inclusions over the signature, and query inseparability, when they give the same answers to existential queries for any instance data over the signature. We demonstrate that different types of corresponding minimal modules for these inseparability relations can be automatically extracted from large-scale DL-Lite ontologies by composing the tractable syntactic locality-based module extraction algorithm with intractable extraction algorithms using the  multi-engine QBF solver AQME. The extracted minimal modules are compared with those obtained using non-logic-based approaches.

Full Text: PDF